Sales Director & Team Lead

Flexcompute and Northrop Grumman reduce space mission preparation time by 100X using AI Physics models, enabled by NVIDIA. PR Newswire Today at 12:30pm PDT The AI Physics breakthrough will expedite product development and support advanced space missions and robotics BOSTON, April 21, 2026 /PRNewswire/ - Flexcompute, the physics company, and Northrop Grumman have developed a foundational AI infrastructure powered by NVIDIA technology that fully automates a simulation workflow to accurately predict thruster impingement effects during space docking in real time. This AI Physics model features built-in uncertainty estimation and is designed to address some of the most complex spacecraft plume interaction challenges in space operations. The work represents a major advancement in how AI Physics models are trained and validated for mission-critical decision-making. Accurate spacecraft control has traditionally required months of mission preparation, driven by the need to generate massive datasets through high-fidelity physics simulation. A central challenge is plume impingement: the physical interaction between a rocket thruster's exhaust plume and nearby spacecraft structures. Simulation is essential to understanding these interactions. In the vacuum of space, gases expand rapidly, creating complex forces and thermal effects that are extremely difficult to reproduce on the ground. With conventional approaches, models trained by these large simulation datasets often need millions of simulations to achieve acceptable coverage and reliability. Flexcompute's AI Physics approach fundamentally changes this process. By training models that learn efficiently from physics-informed structure and data, Flexcompute and Northrop Grumman can make accurate predictions in seconds rather than months while providing explicit uncertainty estimates required for robust control. This could reduce mission preparation timelines by up to 100X. This innovation is built on NVIDIA Physics NeMo, an open-source framework for AI Physics models, which Flexcompute extended with customized model architectures, physics-aware constraints, and training strategies tailored to complex nozzle plume impingement and space robotics interactions. By tightly coupling AI Physics with advanced physics simulation and incorporating uncertainty estimation at inference time, Flexcompute has created a model that delivers fast, reliable predictions engineers can trust for mission-critical control decisions. "At Northrop Grumman, we're pioneering physics AI to accelerate design and solve complex simulation and modeling problems like plume impingement - critical for station keeping, rendezvous, and space robotics. Simply put: we're pushing the boundaries of advanced space operations." said Fahad Khan, Director of AI Foundations, Northrop Grumman. "Partnering with Flexcompute and NVIDIA, we're accelerating innovation and mission timelines to deliver superior space capabilities for customers at the speed they need." By compressing mission preparation cycles and enabling uncertainty-aware predictions, the work also supports more efficient spacecraft operations. Improved plume interaction modeling allows for more confident control strategies, lighter structural margins, and more efficient fuel usage, contributing to longer mission lifetimes and more sustainable space operations. "Northrop Grumman's confidence reflects what sets Flexcompute apart," said Vera Yang, President and Co-Founder of Flexcompute. "We are able to take the most accurate and scalable physics foundations and evolve them into highly trained, customized Physics AI solutions that engineers can rely on. This work shows how we are transforming the role of simulation, not just speeding it up, but expanding what engineers can confidently solve and how quickly they can act." "The industry's most ambitious space missions now demand a level of speed and precision that traditional engineering cycles can no longer sustain," said Tim Costa, vice president and general manager of computational engineering at NVIDIA. "By integrating NVIDIA PhysicsNeMo, Northrop Grumman and Flexcompute are transforming complex simulations like plume impingement from days of compute into seconds of insight, drastically accelerating the path from mission concept to orbit." This work reinforces Flexcompute's role as a trusted technology partner for organizations tackling the world's most complex physics challenges. By combining highly accurate and fast physics simulation with the ability to develop deeply customized AI Physics models, Flexcompute is advancing beyond traditional tools and building the superintelligence platform for physics simulation. The result is a new foundation for how physics, AI, and engineering intelligence come together to shape the future of aerospace, robotics, and beyond. Media Contact SOURCE FlexCompute Inc. This is a paid placement. For further inquiries, please contact PR Newswire directly.

Flexcompute brings photonic simulation to Cadence. New connector enables GPU-accelerated multiphysics modeling and seamless design-to-simulation workflows for AI-driven silicon photonics. Flexcompute announced the launch of the PhotonForge Connector for Cadence Virtuoso Studio, integrating its GPU-accelerated photonic simulation technology into the Cadence Electronic-Photonic Design Automation (EPDA) ecosystem. The connector enables engineers to move seamlessly between photonic layout and high-fidelity multiphysics simulation within a unified two-way workflow. As AI computing continues to scale, data centers are facing growing communication bottlenecks. Electrical interconnects struggle to keep pace with the bandwidth and efficiency demands of large-scale workloads, driving the adoption of silicon photonics as a critical technology. Designing these photonic devices requires accurate modeling of complex electromagnetic, thermal, and electrical interactions. Yet traditional workflows often require engineers to reconstruct device layouts inside simulation environments, slowing iteration cycles and introducing inconsistencies between design and analysis. This integration removes this gap. Engineers can automatically convert device layouts into simulation-ready 3D models, run GPU-accelerated FDTD, mode, charge, and heat simulations, and return validated results to their Cadence design environment, all with native foundry process design kit (PDK) support. This two-way workflow enables teams to explore design variations faster, run parametric sweeps and optimization, and compute S-matrices while maintaining the physical accuracy required for photonic devices used in AI data centers. Flexcompute's photonic simulation technology is already used by leading photonics foundries and research institutions, including GlobalFoundries, Imec, AIM Photonics, and Fraunhofer HHI. This development reflects Flexcompute's broader strategy to build an open engineering ecosystem where leading design platforms and advanced physics simulation work together. By collaborating with industry leaders like Cadence, Flexcompute is helping innovative companies push the boundaries of photonic device design and accelerate the technologies powering AI data center infrastructure.

Flexcompute has launched the PhotonForge Connector for Cadence Virtuoso Studio, integrating GPU-accelerated photonic simulation into Cadence's Electronic-Photonic Design Automation ecosystem. The connector enables engineers to move between photonic layout and multiphysics simulation within a unified two-way workflow. The integration addresses communication bottlenecks in AI data centers, where electrical interconnects struggle to meet bandwidth demands, driving adoption of silicon photonics. Engineers can automatically convert device layouts into simulation-ready 3D models, run GPU-accelerated simulations, and return validated results to their Cadence design environment with native foundry process design kit support. Flexcompute's technology is used by GlobalFoundries, Imec, AIM Photonics and Fraunhofer HHI. The development supports the industry's investment in co-packaged optics for AI infrastructure.